Optimized fringe removal algorithm for absorption images

Optical absorption imaging is a basic detection technique for obtaining information from matter waves, in which the absorption signal can be obtained by comparing the recorded detection light field with the light field in the presence of absorption, thereby giving the spatial distribution of the ato...

Full description

Saved in:
Bibliographic Details
Published in:Applied physics letters 2018-10, Vol.113 (14)
Main Authors: Niu, Linxiao, Guo, Xinxin, Zhan, Yuan, Chen, Xuzong, Liu, W. M., Zhou, Xiaoji
Format: Article
Language:English
Subjects:
Absorption
Algorithms
Applied physics
Atomic properties
Image detection
Matter waves
Noise
Principal components analysis
Spatial distribution
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Optical absorption imaging is a basic detection technique for obtaining information from matter waves, in which the absorption signal can be obtained by comparing the recorded detection light field with the light field in the presence of absorption, thereby giving the spatial distribution of the atoms. The noise in detection arises mainly from differences between the two recorded light field distributions, which is difficult to avoid in experiments. In this work, we present an optimized fringe removal algorithm, developing a method to generate an ideal reference light field, avoiding the noise generated by the light field difference, and suppressing the noise signal to the theoretical limit. Using principal component analysis, we explore the optimal calculation area and how to remove noise information from the basis to allow optimal performance and speed. As an example, we consider scattering atomic peaks with a small number of atoms in a triangular lattice. Compared with the conventional processing method, our algorithm can reduce the measured atomic temperature variance by more than three times, giving a more reliable result.
ISSN:0003-6951
1077-3118
DOI:10.1063/1.5040669